Vm-dev August 2011

vm-dev@lists.squeakfoundation.org

40 participants
111 discussions

ffi-test-main.c fails on FreeBSD
by Gary Dunn 12 Aug '16

12 Aug '16

In conjunction with getting Open Cobalt running on FreeBSD I have been looking into the problems with SqueakFFIPrims on FreeBSD and have reached a point that is beyond my understanding. I am not sure if the fix should come from the VM side or the FreeBSD port maintainer. I needed to get source code back into the ports tree so I did a make in /usr/ports/lang/squeak. I already had the 3.9 tarball, so that just repopulated the source and did the initial build. In the "files" folder I found patch files typical for a FreeBSD port. These patch source files from a generic tarball to work with FreeBSD. The one of interest was patch-platforms__unix__plugins__SqueakFFIPrims__ffi-config which contains ----x-----x----- --- platforms/unix/plugins/SqueakFFIPrims/ffi-config.org Wed Apr 26 20:27:53 2006 +++ platforms/unix/plugins/SqueakFFIPrims/ffi-config Wed Apr 26 20:29:00 2006 @@ -39,6 +39,7 @@ case ${abi} in linux) abi=sysv;; + freebsd*) abi=sysv;; darwin*) abi=darwin;; *) abi=libffi; lib="-lffi";; esac -----x-----x----- I then had the folder work/Squeak-3.9-7/platforms/unix/plugins/SqueakFFIPrims In it is a file 00README which explains how to test FFI. It says to run ffi-test-config but I have no such file. It says to type "make" to build a test suite but the build failed. It says to try with make CPU=any ABI=libffi LIB=-lffi which also failed. I traced the problem to the gcc search paths for includes and libraries -- turns out FreeBSD only searches /usr/include and /usr/lib when the required files are in /usr/local/include and /usr/local/lib. After some fiddling I finally had a "main" to test. It failed: ffi-test-main.c failed at line 361. Here is a bit from that file, ending with line 361: ffiInitialize(); for (ul= 0; ul < 15; ++ul) ffiPushSingleFloat(fa[ul]); GO(FFITypeSingleFloat, many); f= ffiReturnFloatValue(); ffiCleanup(); CHECK(f - ff < FLT_EPSILON); I worked out the CHECK macro but could not find what FLT_EPSILON is. At that point I decided to stop and come to you for help. Where do I go from here? -- Gary Dunn, Honolulu osp(a)aloha.com http://openslate.net/ http://e9erust.blogspot.com/ Sent from Slate001

5 8

Squeak VM port to Google Native Client
by Yoshiki Ohshima 19 Feb '12

19 Feb '12

Hello, I'm playing with Google Native Client a bit and made a port of Squeak VM to it. It is far from complete (even the state where backspace or Cmd-p etc. does not work, sound does not work, etc.), It now reached to the state where I can say the basic part of it is working. The source code is available at: http://github.com/yoshikiohshima/NaClSqueak . Also, f you have Google Chrome 11 beta available, try to turn on the Native Client on (by typing "about:flags" into the title bar) and access: http://tinlizzie.org/~ohshima/squeak/squeak.html (The first try may take some time.) - It is gnuify-ed but we're still learning the JIT situation of NaCl. So, it' need some work to port Cog. - The image data is embedded as a C array in this executable. You can remove EMBEDDED_IMAGE_FILE flag off, and comment/uncomment the corresponding part of squeak.html, it can load the image separately via XMLHttpRequest of the browser. - There is some strange issue when you do fast drag the system window. - As you can see, the source code is derived from the Unix VM. It was hard for me to write some CMake stuff to set it up for the cross compilation. A help to merge it back to the Unix code would be greatly appreciated. - If Google takes over the world and all apps become a Native Client (may be even other browsers^^;), this port may be useful^^; -- Yoshiki

9 25

Cross compile Window from Unix/Mac os
by Jean Baptiste Arnaud 14 Dec '11

14 Dec '11

Since the last Igor ( thank ) update it 's possible to cross compile windows VM(last version) from MacOs and Unix. Using Mingw and cmake. Now, i do for Cog, and for ARM9 processor. My process imply to fix some window file cause some include are wrong, then it the cross compiler don't find some include beginning by Uppercase(cause the file isn't beginning by uppercase). In /Platform/win32/vm/sqWin32Intel.c #include <float.h> instead #include <Float.h> in /Platform/win32/plugin/SocketPlugin/sqWin32NewNet.c #include <ws2tcpip.h> instead #include <Ws2tcpip.h> Regard Jean Baptiste Arnaud jbaptiste.arnaud(a)gmail.com

4 8

Direct object pointers vs indirect ones pros and cons
by Igor Stasenko 26 Oct '11

26 Oct '11

Hi , all RoarVM using object tables, while Squeak VM using direct pointers for object. This is a basic element of VM design, and i wonder, how much impact it does on overall VM speed. Both variants having own advantages/disadvantages, while i think, that with good JIT an extra indirection could be almost insignificant. But having indirect pointers to objects opens quite sexy perspectives. Being able to freely choose object(s) location means that: - its quite easy to implement object memory paging (swapping between memory and disc), - place particular object(s) into special memory location (good for FFI, object pinning etc) - #become: is O(1), instead of O(heap size) The downside of indirect pointers is, of course, a higher memory traffic, which directly impacts all operations everywhere. What else? I'd like to know, what you thinking about it, and why Squeak VM, in particular, using direct object pointers? What are this choice based on? I'd like to know. Maybe i missing something. -- Best regards, Igor Stasenko AKA sig.

10 23

Re: [Pharo-project] jenkins cog - slow monticello on osx
by Igor Stasenko 25 Sep '11

25 Sep '11

On 26 August 2011 13:30, Tudor Girba <tudor(a)tudorgirba.com> wrote: > Hi, > > Since a couple of weeks I am working with the Cog VM built on Jenkins: > https://ci.lille.inria.fr/pharo/view/Cog/job/Cog-Mac-Cocoa/ > > Here is a little feedback. It worked quite well, the only problem is that Monticello is really slow in the Moose image. This means that even opening the Monticello Browser takes several seconds. > > I thought that this might be induced by something in the Moose image, but if I try with the Elliot's VM (2489), it works blazingly fast. > Thanks for the information, Tudor. Apparently the difference between Eliot's VM and Cog-Mac-Cocoa that they are using different frameworks (Carbon vs Cocoa) and i don't know too much details about the difference there. John or Esteban may have better insights how to fix that. > Cheers, > Doru > > > -- > www.tudorgirba.com > > "The coherence of a trip is given by the clearness of the goal." > > > > > > -- Best regards, Igor Stasenko AKA sig.

5 8

About sqSCCVersion.h
by Igor Stasenko 13 Sep '11

13 Sep '11

Hello, i was looking how to make same trick for git, to embed the revision information into the source and found no easy solution. Using gitattributes is powerful, but too tedious for me to set it up fast and too much info for doing simple things. (You may look http://progit.org/book/ch7-2.html) So, to make it fast, i will use shell script to generate output. Unfortunately, it makes little sense to use #define GIT 1 (in sqSCCVersion.h) because its easier to just include generated file: #include "vmVersionInfo.h" // defines REVISION_STRING static char *sourceVersionString() { return REVISION_STRING; } Where vmVersionInfo.h contents is generated dynamically by shell script. To extract an URL a following command does it well: git config --get remote.origin.url Which shows: git@gitorious.org:cogvm/blessed.git To extract commit info, i using: git show HEAD --pretty="Commit: %H Date: %cd By: %cn <%cE>" Which shows: b0df6dc6902a9eb05de02782a77712d855d0422e Date: Wed Jul 20 16:22:16 2011 +0200 By: Igor Stasenko <siguctua(a)gmail.com> So, i will create a script which will generate a header file with these two lines concatenated, and then just include it into modified sqSCCVersion.h The problem is that i cannot put invocation of this script it into cmake config files, because then it will work only if you check out source from git , but not for tarballs. So, to generate the version info file this command will run separately on server. The output of sourceVersionString() function will be something like that: git@gitorious.org:cogvm/blessed.git Commit: b0df6dc6902a9eb05de02782a77712d855d0422e Date: Wed Jul 20 16:22:16 2011 +0200 By: Igor Stasenko <siguctua(a)gmail.com> If you have better suggestions how this string should look like (format, fields) i am here to listen. :) -- Best regards, Igor Stasenko AKA sig.

3 4

About VMConstruction-Plugins-OSProcessPlugin-dtl.27
by Igor Stasenko 13 Sep '11

13 Sep '11

Hello, David is it possible to merge with VMConstruction-Plugins-OSProcessPlugin.oscog-eem.31 ? Since for Cog i using this version. I guess there is no functional changes, only pruning/cleaning unnecessary stuff, but i want to know for sure. -- Best regards, Igor Stasenko AKA sig.

2 3

Fwd: [Pharo-project] Squeaksource.com (was: [squeak-dev] Re: Addition to "About Pharo" dialog)
by Alexander Lazarević 05 Sep '11

05 Sep '11

When it looked like most of global open source Smalltalk development came to a halt last weekend because of squeaksource.com being down for about two days, I asked about moving vital parts like vmmaker to source.squeak.org . This would take some load from squeaksource.com (even if it is until there is a better alternative) and could make commits to vmmaker faster. As suggested I move the discussion to vm-dev. Alex ---------- Forwarded message ---------- From: Bert Freudenberg <bert(a)freudenbergs.de> Date: 2011/8/27 Subject: [Pharo-project] Squeaksource.com (was: [squeak-dev] Re: Addition to "About Pharo" dialog) To: The general-purpose Squeak developers list < squeak-dev(a)lists.squeakfoundation.org> Cc: Pharo-project(a)lists.gforge.inria.fr On 26.08.2011, at 18:14, Alexander Lazarević wrote: Wouldn't it make also good sense to move some vital stuff like the vmmaker repo to this destination? I know this will be controversial, but the current situation is bad at best. That would be a good idea IMHO - you might ask what the other VM developers think of the idea on the vm-dev list.

9 12

what's the standard way of finding out from Smalltalk if one is running headless?
by Eliot Miranda 31 Aug '11

31 Aug '11

I thought to test things like Display extent & DisplayScreen actualScreenDepth but these either fall back to values in the pre-headless image or don't fail and answer the actual display depth, even though the VM is running headless (at least on Mac they do). -- best, Eliot

3 2

Class format
by Michael van der Gulik 31 Aug '11

31 Aug '11

Hi. Are there really 14 bits free in the Class "format" field? The comment for Class>>format: -- http://gulik.pbwiki.com/

3 4

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

Vm-dev August 2011